Hints for: convert vse strip to sound object #3
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Very cool project.
If you want here are some snippets:
Convert audio strips in the vse to sound objects: https://gist.github.com/tin2tin/9b24dff1ab4294e9562724614e450599
Process selected strips: https://blender.stackexchange.com/a/154436/37272
Add-on for installing python modules: https://github.com/amb/blender_pip
These are good snippets. Maybe you know, how can you add strips to a sound object's NLE track with bpy? I can't seem to figure it out.
If you mean for the NLA, I never used it, so I don't know. Aren't Speakers added automatically?
If you mean for the VSE, then Scene strips might be a solution(but I remember some bug-reports on sound). The annoying thing about Scene strips is that you can't reference the scene you're in currently, so you'll have to reference the scene from another scene. Downside is that you can't adjust the volume or anything else sound related in the VSE: https://developer.blender.org/D10029
The surround sound of the VSE is pretty poorly implemented: https://developer.blender.org/T47140
Looking at the VSE pan settings, https://developer.blender.org/T69684#784634 I wish we had implemented the preset buttons back then. I have no idea what these values mean now, why there are 3 columns, and why the same numbers are repeated, and I have no surround setup(are there any free emulators out there?).
The coding of the preset buttons should be very quick to do - the review process, properly much longer, but at least the devs will know there is a problem there.
If you have time, could you take a look and see how they're making sense? Maybe the columns are connected to different number of channels?
Could you check if these values are correct in the vse?
Mono
No pan
Stereo
Left = -1
Center = 0
Right = 1
4 Channels
Left = -1
Center = 0
Right = 1
Rear Left = -2 ?
Rear Right = 2 ?
5.1 Channels https://www.youtube.com/watch?v=RHbtFnjFcp0
Left = -1
Center = 0
Right = 1
Side Left = -1.250
Side Right = 1.250
Rear Left = -2 Center?
Rear Right = 2 Center?
7.1 Channels https://developer.blender.org/T69684#784634
Left = -0.33335
Center = 0
Right = 0.33335
Side Left = -1.66667
Side Right = 1.66667
Rear Left = -1.222
Rear Right = 1.222
I ended up spending the day trying to set up the Surround presets for the VSE in the UI, however I still need the right values for the positions: https://github.com/tin2tin/surround_preset_ui
So I've never used the VSE in Blender, if I need to do video I usually just use Resolve but it's interesting.
I'm not sure how the sound settings work in the VSE, what I started doing in this project is creating arrays of sound sources (or static particle systems), converting them to speakers, and then rendering them out as 5.1s or 7.1s, which seemed to work okay. I then added the ADM object rendering features which just renders groups of non-overlapping speakers to mono, and then takes the speaker motion animation and converts it to ADM sound object panning.
What are these for? Are they coefficients for something?
https://github.com/iluvcapra/soundobjects_blender_addon/issues/3#issuecomment-878876806
What I'd been trying to do is making it so a single speaker can play multiple sounds one after the other with the NLA track but I can't figure out how to script this, the py API in this area doesn't seem to follow the underlying data model here.
Looking at your mono panning cheat sheet...
So the "pan" parameter in the VSE is like an azimuth? And it seems to be normalized so 2.0 is directly behind you?
It looks like column 1 on the cheat sheet is the settings if you are panning in 7.1 render, column 2 is if you are panning into a stereo render, and column 3 is if you are panning into a 5.1 render, you use these values in order to obtain a downmix without attenuation, because I guess the way Blender renders surround sound in OpenAL there might be dead zones or dips between the speakers, because it's 3D rendered and they don't use an equal-power pan law? Just speculating.
I watched your video looks like magic to me! :-)
I don't know much about sound or surround at all, I just annoys me that the UI is so incomprehensible, but I have no hardware to check the values on.
In the VSE, you import(Add Menu > Sound) an audio file, select in and in the sidebar, check Mono(only mono) can be panned, and then compare the values I found here with where you hear the sound coming from in the various pan values in teh sidebar and channel options found in the Preferences: https://github.com/iluvcapra/soundobjects_blender_addon/issues/3#issuecomment-878876806
Then I can add those values to the patch and submit it, and hopefully it'll help some users out there, if the bf devs commits it.
I got a reply from the coder of the lib Blender uses for audio, so everything should be okay now: https://github.com/neXyon/audaspace/issues/34#event-5016013368
I read that, sounds like I suspected. It looks like they use the ITU BS.2094 angles mostly for the surround formats.
https://www.itu.int/dms_pubrec/itu-r/rec/bs/R-REC-BS.2094-1-201706-I!!PDF-E.pdf